Just noting another example of dickhead abuse of the anonymous moderation misfeature of slashdot. Today's case in point is a couple of scattered negative "troll" moderations that were obviously targeted at my recent posts for no good reason, though the obvious bad reason was to anonymously attack my karma and reduce the visibility of my "annoying" posts if I lost my karma bonus.

In one sense, I don't mind having /. "foes". Looking at some of the posts, there are some pretty humongous fools around here, and I don't at all mind recognition as an adversary of such. What I do mind is their covert censorship strategies by gaming the system, and their use of anonymity to avoid accountability for their actions.

As I've frequently noted, there are legitimate uses for anonymity. However, in general such legitimate cases are few and far between, and I've never seen an example of a legitimate need for /. anonymity on moderation. Abusers of the moderation system should obviously be recognized as such--and have their moderation points reduced accordingly.

Also noting another CSS formatting bug. Apart from the problems with Opera, this time it isn't even formatting <p> correctly in IE. I tried adding a </p> tag, but it didn't help.

I'm increasingly convinced that /. is a mostly a waste of time, but I'm still considering if there are any metrics of "effective participation" here. Right now, I think friend and foe designations may be the most interesting metric, especially the foe designations from nasty small-minded people. If you're not offending evil people, then you must be doing something wrong. Interestingly enough, since I adopted the following truth-related sig, I have been the target of quite a number of "foe" designations. I have looked at some of the posts of those people, this actually seems to be a situation where /. is not broken as designed, but is actually working pretty well.

Here's the sig that's bending the Busheviks and Rusheviks so far out of shape:

The truth alone will not make you free, but it's a prerequisite.
Only by knowing the truth can you choose freedom.

Unfortunately, I can't reciprocate as much as I'd like to, since I've already used up my 200 "friend" and "foe" designations. That's another area where /. is broken, but I guess it sort of makes sense, since it could consume quite a lot of storage. It would be quite easy to designate thousands of them.

I've used quite a few sigs on /. over the years, though I don't recall any other that has had so many strong and "effective" reactions. I saved some of the sigs in my preferences, so here they are:

Longhorn? But a steer has no balls! It takes REAL balls to claim Microsoft software "just works"! The modern BIG lie.

The thoughtful we write at once, but insightful takes a little longer... Too bad the moderators have already left.

Everyone's crazy save thee and I, and sometimes I wonder about thee.(Always a metamod, never a moderator (but once).

Microsoft presumes I am guilty of "non-registered" Windows after they rammed it down my throat! I refuse to REregister!

Everyone's crazy save thee and I, and sometimes I wonder about thee. OTOH, Microsoft is crazy, greedy, monopolistic,...

The following was a suggestion email message submitted to the sysops of /.:

Sometimes /. is *VERY* frustrating. It has some good points, and a lot of potential, but it could be so much better. I'd actually offer to help pay for some of the obvious improvements, but looking at it as a business model, my conclusion is that I'd have to be crazy to invest much money in your system. I am looking for better investments than Euros and Japanese real estate, but /. doesn't seem to be a candidate, in spite of its potential...

Anyway, on to the suggestions. Two of these are fairly minor items, but thinking about them gave me an idea for a more substantive and profitable suggestion that I'll offer first: You should sell "improvement funding" for /. enhancements. The way it would work is that you would evaluate potential improvements in the code, and the cost of implementing those features. Then you post those ideas and let people contribute towards the costs. When a particular idea reaches the funding level where it is paid for, then you would implement it. Probably best to do it with a new section, like enhance.slashdot.org, with a list of proposed enhancements, public discussion of each one, and the current level of committed funding that people have donated. You don't need to be exact about the cost, but you should be able to estimate roughly how much programming time is involved. The cost of each project should also include the total costs of the evaluation, and that also needs to include the evaluations of projects that never got funded. (I don't know if you have staff meetings for enhancements, but you should, and that's a legitimate business cost.) Also, you should include the costs of testing and other expenses related to improving the system.

As a concrete example of how this would work, I'll describe how this proposal would apply to my second suggestion. This one is only a minor problem, but annoying: first poster stupidity. It just contributes noise. Enthusiastic, rarely witty, but usually stupid. My suggestion is that you put in some timing constraints on the reply function that would inhibit this form of stupidity. The exact constraints are not important, but let's suggest a "short-reply block" on the first post. Any non-subscriber who tries to make an extremely short initial post will be suspected of first-post abuse, and they will be told they have to wait for a few minutes before they can post--by which time some real poster will have had time to make a substantive comment with real value.

You would evaluate this suggestion and might determine that it would take about 5 hours of programmer time at $50/hour, for $250, plus it took 20 minutes to discuss it in your improvements meeting of three people for another man-hour, for a total cost of $300. Let's call the overhead 33%, and you'd evaluate the cost of this improvement as $400. We'll say that in this case you don't need much time for preparing the post because the original submission is adequate to start the discussion in enhance.slashdot.org.

You might group this proposed enhancement with several other suggestions to address the same problem. For example, someone else might have suggested a first-post relevance-based filter that is estimated to take 80 hours of programming time. (It would compare the linked articles to the submitted post to see how likely it was that the poster had actually read the article, and also check for too much cutting-and-pasting.) Readers of /. could look over the ideas, make comments, and even contribute to the funding of any idea. If possible, this should be done with some kind of dynamic meter so people could see how that project was doing for funding. I'm thinking of one of those thermometer things. When the funding is available, then you schedule the implementation. You could even include a priority factor with some extra funding for rush jobs.

(Two more wrinkles should be mentioned:

Wrinkle One: What about projects that don't get enough funding? I think you should explicitly state that you will keep those donations for other improvements. However, all of the people who donated to that project will be notified when it is about to expire, and any of them can request that the discussion be extended. For example, the default funding period might be one month, but someone might want to insist on discussing it for another month. That person would have to add some explanation to the top post of the discussion to explain why he or she thinks the idea is still worth implementing.

Wrinkle Two: What about removing features? I think you should treat them as "code streamlining enhancements" that would be proposed and funded in the same way. However, these would probably usually be prepared internally by /. programmers based on evaluating substantive problems with the actual code.)

Okay, now for the third suggestion, which is related to the continuous problems with moderation. (However, in the funded suggestion scenario, these moderation problems would actually be a source of ongoing funding for /. enhancement, thus converting the problems to something useful.) I think the moderation reports should include the identity of the moderator. This would make it relatively easy to spot moderators who are acting out of personal vindictiveness. I strongly believe that often happens to me, but I'm actually proud of having certain nasty little people as personal enemies. For example, it doesn't bother me at all when some racist religious fanatic is attacking me, but at the same time I don't much like it when they can use your moderation system to do it anonymously.

Background: As a kind of quality control over visitor contributions, /. uses moderation (M1) and metamoderation (M2). M1 rates posts by various criteria, and then M2 is supposed to doublecheck the moderators (M1ers) by asking a broader segment of the /. community to agree or disagree with those M1 decisions.

Problem #1: Lots of bad M1 ratings. Many posts are rated inappropriately, even ridiculously inappropriately. And yet, according to the FAQ and sysops, M2 is 90% in agreement with M1.

Hypothesis: Bad M1ers are gaming the system. (Shall we call them immoderate moderators?) The system says that M1 is linked to M2, but what it *REALLY* means is that aggreeing with existing M1s in your M2 ratings *CAUSES* you to M1 again. It is a closed feedback loop that rewards lying--but it allows the sysops to think that the M1 system is 90% "accurate".

Problem #2: If you use M2 to comment on bad M1s, you will be below 90% "accurate" and you will never get to do M1 again. My estimate is that current mods are less than 70% reasonable.

Basis of this belief: Careful consideration of the English words (AKA epistemological analysis) as applied in this context. I think these are some of the principles that should inform moderation:

Insightful

Insightful also implies accuracy, so a post with a counterfactual statement should not qualify for an insightful moderation. Disagreement over a matter of opinion in a post should not disqualify it from being insightful, but it should go deeper than the surface of the issue. The post should also actually reveal the insight, not simply hint at it. That means one-line insights are rare.

Informative

An informative post should offer more information than a mysterious link, and again, it should not be rated informative if it is wrong. It should also be relevant and on-topic, and preferably new.

Interesting or funny

Mods for interesting and funny are matters of opinion, and in this case the M2er's opinion is just as valid as the M1er's. In general I'll try to round those in the M1er's favor for the inverse reason. For example, if a post is rated funny, but I don't find it funny, then my M2 rating should disagree, even though I might think there are some people who find it funny.

Redundant

Not sure how long I've been using /., but I still don't know what this moderation is really supposed to mean.

Overrated

Blanket negative ranking. I think this one is basically a bad idea. M1 ratings should be clearly paired positive and negative. As it stands, this apparently means "My opinion is that I want fewer people to see this post." I suspect this is a primary mode of M1 abuse, mostly because I believe there is also an "underrated" rating that almost never gets used.

Solution: The sysops should clarify the principles of moderation. They should also check how many of the M1ers are simply gaming the system by always agreeing 100% in their M2 rankings in order to get more chances to M1.

At this point I've started noticing lots of obvious idiots, and I'm just marking them as foes so their annoying posts will be less visible to me. Mostly that translates into mindless Busheviks, but it also includes other categories like aggressive creationists and no-understanding-of-accounting idiots who think space research can be funded out of petty cash in a free market.

That approach also defines the opposite perspectives as candidates for the "friends" list, especially on the Dubya topic.

Anyone who claims they voted for Gore over Dubya in 2000, but is going to vote for Dubya over Kerry in 2004 is either crazy, stupid, gullible, lying, or some combination thereof.

The post where I noticed the sig was rather dumb and useless, which mostly supports the second option.

The other one is, to the best of my understanding, a true goddess of hate speech. I'd call her anti-patriotic and anti-democratic, but that would be a reasoned approach, and her blather transcends reason. Since the actual post I noticed was not pure blather, it is safe to assume the author is not the real person, but just a sicko admirer.

Typical mindless Bushevik sig saying terrorists hate Dubya. Point of fact is that you always want your enemies to have incompetent leadership, and the terrorists can't possibly hope for more incompetence than Dubya provides. The fuel of terrorists is hate, and Dubya has created enough to top off their tanks for MANY years.

Another moronic Dubya supporter. The post where I noticed him was about Gilligan's Island and Monty Python. 'Nuff said.

"Free" broadcast television turned out to be worth even less than we paid for it. An entire generation of wasted minds.

Typical lying Bushevik sig, claiming UBL endorses Kerry. In actuality, hate (like his) is what fuels terrorism. Dubya is doing such a great job of increasing the international hatred against America that UBL has to love Dubya and be praying for his continuation in office. Probably wouldn't endorse Dubya, however. I think UBL would prefer to see another coup like in 2000.

Well, his post was most notable for all the obscenities, but then I noticed his anti-Michael-Moore sig and decided to officially tag him as a foe. He cites one of the more idiotic lists of supposed lies in Fahrenheit 9/11 , except that there aren't any lies in the list, just various half-baked claims that that author's interpretations of certain carefully selected facts (mixed in with some half-truths and outright BushCo agitprop) are better than Michael Moore's interpretation of some other facts.

I've decided to start annotating why I mark someone as a friend or foe, for what little that's worth. In the case of the Old Burke, it's for his signature equating patriotism with voting Republican. That is a completely false equivalence, more suitable to the Stalin's Soviet Union, simply substituting "Communist" for "Republican".

My own observations are that the "modern" GOP voters are mostly in three classes: Greedy rich bastards, single-issue fools, and people who always vote that way.

Statistically, I believe the first group is actually quite small, mostly because rich people are scarce, but also because not all rich people are greedy and nasty. However, this first group is where the GOP money comes from--they don't donate to politicians. They invest in bribable ones like Bush. Dubya's largest contributor = Enron. I rest my case.

The last group is the largest (for both parties), and least "helpful" for real democracy. Though numerous, they are most effectively neutralized by low turnout, because most of them are rather apathetic. Why should they care? They already know which way they're going to vote. My theory here is that the balance is most influenced by conditions when they were growing up. If they grew up poor, as in the Depression or in a poor family, they tended to become Democrats, but if they grew up in a rich family, they are more likely to be Republicans. Historically, the GOP was weaker here, but these years it's shifting in their favor.

The second group, though smaller, is large enough to be crucial for swinging elections through special interest appeals. They are also the ones that can be motivated to deliver large and reliable blocks of votes. These are the fanatics like gun nuts or anti-abortionists. Just appeal to their issue, and they'll deliver their votes. They go beyond being unhelpful to real democracy--they act to destroy it.